How the Column Mapper Works
AnyTool maps source headers to a canonical import schema entirely in the browser. The CSV is parsed with PapaParse, then each source header is normalised (lower-cased, non-alphanumerics collapsed) and scored against a built-in field schema and its synonyms using Jaro-Winkler similarity, with exact and substring matches boosted toward 1.0. Each column gets a best target (left unmapped below a confidence threshold), a confidence score and ranked alternatives. The tool detects when two source columns map to the same target and when a required field is unmapped. You override any mapping from a dropdown, then export the renamed/reordered CSV or a reusable JSON mapping. Nothing is uploaded.
- Runs client-side — the file is never uploaded
- Jaro-Winkler header → canonical-field matching with synonyms
- Confidence scores + ranked suggestions; threshold for auto-map
- Detects duplicate-target conflicts and unmapped required fields
- Export mapped CSV or JSON mapping for reuse
- No CDN, no backend, no signup
